In a conventional data processing apparatus, temporary data, e.g., data temporarily stored in a register of an electronic circuit unit, is erased when an electric power of the apparatus is off and cannot be restored even after a power recovery. Accordingly, in the conventional data processing apparatus, such temporary data may need to be backed-up on an external storage, e.g., a magnetic storage medium, during a power shutdown, and restored after a power reactivation.
However, data backup/restoration using an external storage such as a magnetic storage medium requires additional software support. Further, data backup/restoration increases the number of processing steps to be performed when the power is shut down and reactivated, and thus, lengthens the processing time, which may cause stress to a user.
In view of the above, there are disclosed and suggested a data processing apparatus capable of performing a data backup/restoration at the time of an emergency power shutdown without performing a data backup operation during normal operation, and a data maintenance apparatus capable of being appropriately embedded in a device operated by a power supply voltage.
In the above-mentioned data processing or maintenance apparatus, however, it is sometimes necessary to perform a reconstruction of the logical configuration between the electronic circuit unit and the storage medium, e.g., the addition of scan paths. This makes it difficult to apply the above-mentioned apparatus to existing devices. Another problem exists in that the above-mentioned apparatus is disadvantageous in terms of design and manufacturing costs.
As a solution to these problems, it can be considered to record temporary data on a FeRAM (Ferroelectric Random Access Memory). Since FeRAM is a nonvolatile storage medium, data recorded thereon is maintained even when the power is off. However, FeRAM has a longer access time when compared with an ordinary RAM (Random Access Memory), which deteriorates data processing performance during normal operation.